Defining the Operative Time Threshold for Safety in Patients Undergoing Robotic Pancreaticoduodenectomy

Abstract

Introduction: Robotic pancreaticoduodenectomy (RPD) is a safe and efficacious operation in appropriately selected patients. However, the increased complexity of a robotic approach results in longer operative times. It is unknown at what operative time the risks of a longer RPD exceed those of a shorter open pancreaticoduodenectomy (OPD). Method: The NSQIP database was queried for patients who underwent RPD and OPD between 2014-2019. Emergent operations and vascular resections were excluded. The benchmark OPD control was defined as the fastest tertile with regards to operative time. Multivariable and sensitivity analysis were employed to identify the operative time threshold (OTT). Results: 6,270 patients were included in the study with 939 (15.0%) undergoing RPD and 5,331 (85.0%) as a benchmark OPD control. Incidence of major morbidity or mortality was similar in the RPD and OPD cohorts (32.8% vs. 35.1%, p=0.17). The median operative time for RPD was longer than OPD (6.7 vs. 4.1 hours, p<0.01). Stepwise logistic regression analysis identified operative time as the only variable associated with increased complications in RPD patients (OR: 1.20, 95% CI: 1.12-1.29, p<0.01). The OTT – the time above which RP patients experience greater than the 35.1% OP benchmark complication rate - was identified as 7.7 hours (FIGURE 1). Furthermore, on subgroup analysis of RPD patients, the slowest tertile had more complications than the middle and fastest tertiles (41.5% vs. 32.2% vs. 24.5%, p<0.01). Conclusion: RPD is associated with similar or improved outcomes when the operative time threshold of 7.7 hours is met when compared to OPD.